This dissertation has been 64—1274 microfilmed exactly as received

LAZEN, Alvin Gordon, 1935- CHEMISTRY OF A PSEUDOMONAS GLYCOPEPTIDE DEMONSTRATING Rh0(D) SPECIFICITY.

The Ohio State University, Ph.D., 1963 Bacteriology

University Microfilms, Inc., Ann Arbor, Michigan CHEMISTRY OF A PSEUDOMONAS GLYCOPEPTIDE DEMONSTRATING

Rh0(D) SPECIFICITY

DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By Alvin Gordon Lazen, B.S.

******

The Ohio State University

1963

Approved by

C r - Adviser Department of Microbiology ACKNOWLEDGMENTS

Science is only a part of life. A step forward in science— Just as a step forward in life— depends upon the help, encouragement, and understanding of many people. To those persons who have helped and encouraged me I wish to express my deep-felt gratitude: Dr. C. I. Randles for his patience and restraint in permitting me to find my own way, while being always willing to help; Dr. M. S. Rheins for his

interest and encouragement; Drs. M. C. Dodd, N. J. Bigley,

J. M. Blrkeland and other faculty members who helped me in many ways; George Hrubant who began this study; and to the many graduate students who shared this part of life with me.

Financial support from the Ohio State Research

Foundation and especially from the Muellhaupt Fellowship of

the Graduate School was indispensable and appreciated aid during the course of my research.

I am most grateful to my wife Lyla, who has provided

all— help, encouragement, understanding, patience, support,

and faith.

ii TABLE OF CONTENTS

Page

Acknowledgments...... 11

Tables ...... lv

Illustrations...... vi

Introduction...... 1

Literature Review ...... 2

Materials and M e t h o d s ...... 8

Experimental Results ...... 2H

Discussion...... 55

S u m m a r y ...... 71

Bibliography ...... 73

Autobiography ...... 79

iii TABLES

Table Page

1. Comparison of the Organism Isolated from "Streamers" and Pseudomonas denltrlflcans . . . 12

2. Elemental Analyses of the Pseudomonas Glycopeptide Preparations ...... 24

3. Folin and Ciocalteau Test of the Glycopeptide . . 29

4. Results of Ninhydrin Tests for Amino Nitrogen . . 30

5» Identification of Classes of Sugars by the Value OD 390- OD 421 mu in the Dische Cysteine Test ( 1 9 5 5 ) ...... 37 6. Identification of Specific Aldohexoses by the Ratio OD 535/440 mp. in the Dische Carbazole Test ( 1 9 5 5 ) ...... 39 7• Quantitation of Hexoses Present in the Glycopep­ tide by Means of Colorimetric Tests ...... 40

8. Calculation of Hexoses Present in the Glycopep­ tide as 50:50 Glucose:Galactose ...... 40

9. Glucostat Determination of Glucose Present in the Glycopeptide ...... 41

10. Comparison of Hexosamines to a Spot Appearing in Chromatograms of Hydrolyzed Glycopeptide . . . 48

11. Comparison of Rf Values of Silver Nitrate-Reactive Spots on Paper Chromatograms Using Various Solvent Systems ...... 49

12. Comparison of Rf Values of Ninhydrin-Reactlve Spots on Paper Chromatograms Using Various Solvent Systems ...... 50

iv TABLES— Continued

Table Page

13. Summary of Data on the Chemistry of the Pseudomonas Glycopeptide...... 60

14. Comparison of Elemental Analyses with Calculated Analyses for a CiyH28n2°12 Un i t ...... 63

v ILLUSTRATIONS

Figure Page

1. Platinum Shadow-Cast Electron Photomicrograph of Pseudomonas on collodion and carbon background Xl4,700 ...... 10

2. Platinum Shadow-Cast Electron Photomicrograph of Pseudomonas on collodion and carbon background *34, 400 ...... 11

3. Isolation and Purification of the Pseudomonas Glycopeptide ...... 17

4. Presence of Reducing Sugars During Acid Hydrolysis of the Pseudomonas Glycopeptide . . 27

5. Ultra-Violet Absorption Spectra of the Glyco­ peptide and Known Sugars in the Ikawa and Niemann Test (1949)...... 33 6. Absorption Spectra of the Glycopeptide and Known Sugars in the Dische Cysteine Test (1955) . . . 34

7* Absorption Spectra of the Glycopeptide and Known Sugars in the Dische Classifying Cysteine Test ( 1955 )...... 36 8. Absorption Spectra of the Glycopeptide and Known Sugars in the Dische Carbazole Test (1955) • • 38

9. Absorption Spectra of the Glycopeptide, Glucose and Glucuronic Acid in the Dische Carbazole Test for Hexuronic Acids (1955) ...... 43

10. Paper Chromatogram of Glycopeptide Hydrolyzates and Known Carbohydrates ...... 44

11. Partial Separation of the Glucose and Galactose Components of the Glycopeptide on Paper Chromatograms...... 46

vi ILLUSTRATIONS— Continued

Figure Page

12. Complete Separation of the Glucose and Galactose Components of the Glycopeptide on Paper Chromatograms ...... 47

13. Determination of Reducing Sugars and Ninhydrin- Reactive Material in Samples of Eluate from a DEAE Cellulose Column ...... 54

14. Possible Linkages of Amino Acids to the Carbo­ hydrate in the Glycopeptide...... 65

15* Comparison of Sialic Acid, Muramic Acid and a Proposed Glycine-N-acetyl Glucosamine C o m p o u n d ...... 69

vii INTRODUCTION

In the highly acidic waters flowing from coal mines in Southeastern Ohio was found an unusual biological phenomenon— an organism embedded in a tough slime and able to survive in waters of pH 2-3. The organism was isolated from the slime and studied in this laboratory to determine its general characteristics. Finally attention was directed to the slime material itself. Interest in the slime material was rewarded by the discovery that a preparation from the slime demonstrated serological specificity in the

Rh0(D) human blood-group system. This discovery provided the initiative to determine the chemical composition of the slime material. It was the multiple purpose of this study to examine the organism producing the slime, isolate and purify the serologically active substance derived from the slime, determine its composition and structure and, if possible, relate the findings to an explanation of the ability of the glycopeptide to cross-react with Rh0(D) antibodies.

1 LITERATURE REVIEW

In 1952* Temple and Koehler reported an unusual type of "streamer" growth In the acid effluent water from West

Virginia coal mines. They described the streamer as fol­ lows : "it exists there as cream colored masses of filaments which are attached to any projection and form very long wavy streamers. These streamers are composed of bacteria con­ tained in a tough slime. . . . A bacterium believed to be the one responsible for the streamers has been Isolated. In stationary media it forms a mucoid growth and in flowing media it forms streamers."

Randles (1957) isolated an organism from mucoid streamer growth in acid mine-drainage water in Southern

Ohio. Hrubant and Randles (1958) identified the Ohio streamer organism as a species of Pseudomonas and initiated nutritional studies directed toward determining the source of nutrients for the organism In its acid water environ­ ment. Prom these studies came the discovery that the bacterium required thiamine for growth. It was postulated that the thiamine might be derived from the thiobacilli also present In the acidic water.

Demos and Randles (1959) continued the nutritional studies and investigated the conditions for optimum slime production. Results showed little specificity on the part

of the Pseudomonas for either carbon or nitrogen source.

The organism grew rapidly and produced slime material in

profusion on media containing a number of simple

monosaccharides and amino acids.

In the course of the above studies it was discovered

that the slime material produced by the Pseudomonas on

laboratory media shared specificity with human Rh0(D) blood-

group substance. The impetus to test the pseudomonas glyco­

peptide in a serological system with Rh antisera was derived

from the coincidence of the postulation by Hrubant that

neuraminic acid was a component of the "polysaccharide" and

the success achieved by Dodd et al. (i9 6 0) in inhibiting

Rh0(D) antibody with N-acetyl neuraminic acid and neuraminic

acid-containing compounds. Hrubant did not confirm the

presence of neuraminic acid but contributed valuable

insights into the composition of the "polysaccharide" in the

form of a negative Biuret test and positive reducing sugar

tests. Dodd's work showed that the slime material was able

to inhibit Rhc(D) antibody but not anti-Rh0(C), anti-E,

antl-c, or anti-e sera. The "polysaccharide" was also shown

to have Rh0(D) specificity by the production of a positive

skin test when a dilute solution of the purified slime material was intradermally injected into a rabbit passively

sensitized 24 hours earlier with anti-Rh0(D) serum. Later,

Bigley et al. (1963) continued the studies of the glycopeptide and reported that the similarity of specificity

of Rh0(D) antigen and the glycopeptide was demonstrable by

the following: "(1) inhibition of anti-D agglutinins by

Pseudomonas glycopeptide; (2) anti-Pseudomonas glycopeptide

rabbit sera contained both saline agglutinins and incomplete

antibodies specific only for cells containing D antigen;

(3) D positive cells absorbed specific agglutinins from

antl-Pseudomonas glycopeptide sera; (4) anti-D serum blocked agglutination of D cells by anti-Pseudomonas glycopeptide serum."

Attempts to associate Rh specificity with other bac­

terial species have resulted in mixed findings. Springer et al. (1961) failed to demonstrate Rh specificity in a

large number of bacterial species and under their test condi­

tions could not Inhibit anti-Rh0 agglutinins with various sialic acid compounds. Bigley et al. (1963) reported no

RhQ(D) specificity in a capsular material from a strain of

Pseudomonas aeruginosa known to contain mannan, RNA, and

DNA. On the other hand, Boyd and Reeves (1961) tested an

N-acetyl neuraminic acid polymer, colominic acid, from a mutant of Escherichia coli, K235 L+0“ which inhibited

Rh0(D) antibody.

Demonstration of blood-group activity other than Rh activity by bacterial preparations is not unusual. Springer et al. (1 9 6 1) review many cases of A,B, and 0 (H) blood- group activity by Gram-negative organisms dating back to the

first unequivocal instance reported by Schiff (193*0 for one

strain of Shigella shigae. However, the pseudomonas glycopeptide uniquely or in partnership with a very few

other microbiological preparations exhibits Rh0 specificity.

It stands as the first reported instance of this specificity by bacterial preparations.

Chemical analyses of bacterial surface antigens dates back to the pioneer work of Heidelberger, Goebel and Avery

(1925) on the capsular polysaccharide of pneumococci. These workers continued their observations in succeeding years until now it is impossible to list all their work in a short review. A book by White (1938) covers the early work and

Heidelberger (1956, 1959) reports many of the later findings.

Once the field was opened reports of various immuno-

logically active surface substances became commonplace. In

1937 Ivanovics and Bruckner discovered the first glutamyl polypeptide capsule while studying Bacillus anthracis and other members of the genus Bacillus. Shortly thereafter

Boivin and Mesrobeanu in a long series of papers (1935* 1936,

1 9 3 7* 1938) introduced the important concept of the presence in Gram-negative bacteria of complex somatic antigens com­ posed of polysaccharide-protein-lipld. So intensive have been the studies on some antigens that research on just a single organism has spanned many years, e.g., Morgan*s work °n Shigella dysenterlae first reported in 1937 and still reaching print 20 years later (Morgan 1937, 19*10; Davies,

Morgan et al., 1954, 1955)* So extensive is the range of organisms studied in the laboratory that it is possible only to cite reviews rather than individual work. Morgan reviewed much of the early work in 1944. Evans and Hibbert (1946) and Stacey (1946) summarized the work on bacterial polysaccharides and the chemistry of mucopolysaccharides and mucoprotelns respectively. The latter review comments on several bacterial polysaccharides which fit into the category of mucopolysaccharides. Neely (i9 6 0) wrote a full account of dextran research while Davies (i9 6 0) in the same volume wrote an excellent review of many facets of the polysac­ charides of Oram-negative organisms. Westphal et al. (1 9 5 8) have published a chapter on the mucopolysaccharides of

Gram-negative organisms in a book worth inspecting for further information on the chemistry and biological signifi­ cance of mucopolysaccharides. Wilkinson (1958) emphasized extracellular polysaccharides of bacteria but treats these in detail with regard to metabolism as well as chemistry and antigenicity.

A recurrent feature of the work on immunogenic sub­ stances has been the reporting of new or unusual mono­ saccharide units and polysaccharides. Whereas early papers established the presence of the common aldohexoses and 7 aldopentoses and noted the Importance of 2-deoxy-2-amino

hexoses and hexuronic acids, the introduction of newer

techniques of analysis brought about reports of increasingly

complex structural units. A class of sugars isolated from

Salmonellae were named abequose and tyvelose, and, mainly

due to the investigations of Westphal et al. (1954), were

found to be 3,6-dideoxyhexoses. Other workers have more

recently reported other 3,6-dideoxyhexoses as well as

3,6-dideoxyhexitols. Webster et al. (1952) found that the

"Vi" antigen is composed of an aminodeoxyhexuronlc acid.

Aldoheptoses have been found in various Enterobacteriaceae and the list of 2-deoxy-2-amino sugars has been lengthened by the discovery of previously unknown amino sugars such as

2-amino-2,6-fucosamine from Chromatlum violaceum (Crumpton and Davies, 1958). Most germane to our present study was the isolation of a polysialic acid, colominic acid, from

Escherichia coli K235 by Barry and Goebel (1957).

It is with the above reported data as a basis and with the research methods developed over the years as tools that the work to be presented here was initiated. MATERIALS AND METHODS

Organism

The organism employed in this study was originally

Isolated by Randles (1957) from masses of tough, gray,

slimy, "streamer" growing in a small stream of drainage

water from the Todd Mine, Vinton County, in Southeastern

Ohio. Profuse growth of the organism was present in the

stream despite the waterfs low pH and high acidity (pH 2.85;

titratable acidity 2,770 rag/1 as sulfuric acid). Pure cul­

tures isolated from the streamer had been lyophilized and it was from one of the lyophilized samples that stock cultures were grown on yeast-extract dextrose of Sabouraud’s medium.

Stock cultures were grown 2k hours at room temperature and stored at 4 C. Transfers were made wery two months. When cultures were desired for use in preparing slime material, a stock culture was transferred every 2k hours on Sabouraud^ agar until the culture rapidly produced a flowing pearl-gray, translucent slime. This generally required only 2 or 3 transfers.

Morphological and biochemical studies revealed that the streamer organism resembled a pseudomonas species.

Electron micrographs prepared by Marie H. Grelder of the Ohio State University Department of Pathology showed the bacillary shape and a single polar flagellum (Figure 1) and the insertion or attachment of the flagellum to the cell

(Figure 2). The data in Table 1 showed that the organism most resembled Pseudomonas denitrlflcans when compared to the description of that organism in Bergey*s Manual of

Determinative Bacteriology (Breed et al,, 1957). In addi­ tion to the reactions reported in Table 1 the organism pro­ duced no acid or gas in xylose, dextrose, inulin, sucrose, inositol, or lactose. It grew well on most laboratory media including brain-heart infusion, trypticase soy broth,and yeast-extract dextrose media. A suitable defined medium devised by Hrubant and Randles (1957) was composed of the following ingredients:

Glucose 40.0 g

Thiamine 500 yig

Glutamic Acid 1.47 g KgHP04 1.0 g

KHgPOn 0.25 g

MgSC>4.7 H20 0.01 g

FeS04*7 H20 0.01 g Agar 15.0 g

Water 1000 ml 10

Figure 1.— Platinum shadow-cast electron photoraicro graph of Pseudomonas on collodion and carbon background X 14,700. 11

Figure 2.— Platinum shadow-cast electron photomicro­ graph of Pseudomonas on collodion and carbon background showing attachment of flagellum X 3^,400. 12

TABLE 1

COMPARISON OP THE ORGANISM ISOLATED PROM "STREAMERS" AND PSEUDOMONAS DENITRIPICANS

Characteristic Pseudomonas denitrificans Streamer Isolate

size and rods 0.5-0.7 by 1 .2 5 M« rods 1.1 by 2.4- arrangement occurring singly and In 2.6|x- same arrange large slimy masses ment motility motile motile Gram stain Gram-negative Gram-negative Gelatin small, circular, contour­ same colonies ed, raised, moist, pearly gray glistening Gelatin stab whitish, lobed surface grayish, growth at growth yellowish green top, no liquefac­ growth in stab, no lique­ tion faction agar colonies pearly white, circular, same entire agar slant broad, whitish, con­ same toured, moist, entire broth turbid, with thick turbid especially wrinkled pellicle at top, pellicle which may sink to bottom litmus milk not coagulated not coagulated potato reddish-gray layer no growth indole not produced not produced nitrate reduc­ reduced with production reduced with pro­ tion of nitrogen duction of gas oxygen require­ aerobic, facultative aerobic ment optimum temper­ 25 C 25-28 C ature 13 Isolation and purification of the gXycopeptlde

The techniques employed In Isolating and purifying the glycopeptide were standard and uncomplicated. However, since the mode of preparation may bear upon the resulting serological activity, the method will be presented in some detail.

Stock cultures of the Pseudomonas were serially trans­ ferred as previously described. When the cultures had rapidly produced a loose, viscous slime on Sabouraudfs agar,

8-10 ml of sterile distilled water were added to the tube.

The growth on the heavily inoculated slant was suspended in the water by lightly scraping the slant with an inoculating loop and shaking the suspension. The contents of one such slant were added to five liter Roux bottles containing

400 ml Sabouraud’s agar and distributed over the surface of the agar. The large surface area and the adequate supply of oxygen in Roux bottles contributed to the growth of the

Pseudomonas and subsequent large yields of the glycopeptide.

Incubation of the Roux bottle cultures was at room tempera­ ture for 48-72 hours. At the end of the incubation period the pseudomonas culture formed a uniform, gray, mucoid, glistening coating over the entire surface of the agar.

The surface of the Roux bottle culture was carefully wiped with glass rods In order to remove the slime-enveloped

Pseudomonas without contaminating the preparation with gouged agar. Several 20 ml aliquots of water were sprayed into the Roux bottles to remove more of the mucoid slime.

Finally the growth was placed in a Waring blendor and agi­ tated for 20 minutes. Ice was added to the blendor cup to maintain low temperatures. Treatment for this length of time was found sufficient to divest the organism of its slime coating. In those cases where the cells were found to still possess significant amounts of slime coating, the collected cells were again subjected to agitation in the

Waring blendor.

The viscous solution containing suspended organisms was then centrifuged in an International angle centrifuge for 20 minutes at maximum speed. Most of the organisms were removed from suspension by the centrifugation procedure but much clearer solutions were obtained by employing a second centrifugation in a Servall centrifuge for 10-20 minutes at maximum speed. The resulting glycopeptide solu­ tion was extremely viscous and ranged from a clear solution to a slightly opalescent solution. The cell-free supernatant solution was then precipitated by the addition of cold 95 per cent ethanol to a final concentration of 75 per cent.

Frequently the glycopeptide precipitated at a slow rate.

Addition of small amounts of sodium chloride accelerated the precipitation process. When enough ethanol had been added, a decided opalescence appeared in the solution. Upon standing in the cold the opalescence developed into increas­

ingly larger floccules which eventually settled. The

precipitate was separated by centrifugation in 250 ml

bottles in the International angle centrifuge and then

redissolved in distilled water. The precipitation and

redissolution steps were repeated a second time. When the

twice precipitated material was redissolved it was placed

in Visking tubing and dialyzed against distilled water in

the cold for 48 hours. Dialysis water was constantly agi­

tated by a magnetic stirrer and was changed four times daily.

A final ethanol precipitation of the dialyzate was then per­ formed using cold 95 per cent ethanol to a final concentra­ tion of 85-88 per cent. The precipitate which now developed was in the form of large white or gray floccules which could then be collected by centrifugation and spread to dry on a six inch petri dish in an evacuated dessicator.

Upon drying the resultant material ranged in color from a light tan to a decidedly brownish color depending upon the thickness of the dried layer. The dried glycopeptide was then pulverized with a pestle, weighed, and stored in a small test tube in an evacuated dessicator.

Average yield from one Roux bottle culture ranged from 0.25 to 0.5 gram. The powdered glycopeptide ranged from light tan to very light gray in some preparations. 16 Preparations made in the above manner were free of

agar contamination as indicated by a negative carbazole test

for hexuronic acid (Dische, 1 9 5 0). Furthermore, centrifuga­

tion and later steps effectively removed pseudomonas cells

since Biuret tests for protein were negative.

Each preparation was tested for specificity in a

precipitin ring test with glycopeptide solution carefully

layered over suitable commercial RhQ(D) antiserum.

A schematic presentation of the steps taken in pre­

paring the glycopeptide is given in Figure 3.

Chemical tests

In the early part of this study an intensive effort was made to apply colorimetric techniques to the task of elucidating the chemical composition of the pseudomonas glycopeptide. Using Dische (1955) as a guide, the Indole,

tryptophane, cysteine, and anthrone tests were used for the quantitative determination of hexoses. It is important to note that the anthrone test in the above reference inaccur­ ately calls for a 2 per cent rather than a 0.2 per cent anthrone in concentrated sulfuric acid solution. At first glucose alone was used to prepare standard curves for hexose determination. Later, galactose was discovered In the polymer and standard galactose curves were also prepared.

The carbazole test (Seibert and Alno, 19^6) was applied In an effort to identify the specific aldohexoses present in the 17 Pseudomonas grown on Sabouraud slant for 24-48 hours at room temperature

i 8-10 ml water added to slant culture and growth washed into Roux bottles containing Sabouraud's medium i incubate at room temperature 48-72 hours I scrape and wash growth off Roux bottle cultures I ------j ------place in Waring blendor and agitate 20 minutes while cooling with added ice J centrifuge in 250 ml bottles i decant supernatant fluid and discard or recycle cells — 1 centrifuge in 20 ml tubes I — decant supernatant fluid and discard or recycle cells J precipitate supernatant with ethanol J collect precipitate by centrifugation I redissolve in distilled water 1 repeat above two steps 1 dialyze i Precipitate with ethanol i spread to dry in petri dish in dessicator 1 pulverize dried material, weigh, and store in dessicator Figure 3.— Isolation and purification of the Pseudomonas glycopeptide. 18 glycopeptlde. Tests for 6-deoxyhexoses were by the Dlsche

and Shettles (19*18) method using rhamnose as the standard.

Tests for hexuronic acids were by the Dische procedures

(19*16 a, b, 1 9 5 0) employing glucuronic acid as the standard.

The Dische (19*19) cysteine-sulfuric acid method was used to

test for the presence of pentoses and heptoses against

standards of arabinose and sedopheptulose. Hexosamines were assayed by the indole-hydrochloric acid method of Dische and

Borenfreund (1950) and the Blix modification of the Elson and Morgan test as described by Gardell (1958). A paper chromatographic method for hexosamine identification described by Gardell (1958) was also employed using the capillary technique of Stoffyn and Jeanloz (195*0 for the ninhydrin degradation step. Glucosamine was the standard in the colorimetric tests and both glucosamine and galactosamine were used as standards for the chromatographic method. The

Reissig et al. (1955) and the Aminoff et al. (1952) methods were used to determine N-acetyl hexosamines using N-acetyl glucosamine as the standard. Tests attempting to determine the presence of sialic acids were performed by the Werner and Odin modification of the direct Ehrlich test described by Gottschalk (i9 6 0) and by the Aminoff (1 9 6 1) method.

Attempts to degrade the glycopeptlde with neuraminidase employed Vibrio cholera neuraminidase obtained from General

Biocheraicals, Chagrin Palls, Ohio. Enzymatic quantitation 19 of glucose was performed with the Glucostat kit of the

Worthington Biochemical Corporation, Freehold, New Jersey.

The Ikawa and Niemann test (19^9) based on the dif­

ferential ability of sugars treated with sulfuric acid to

absorb ultra-violet light was employed in an attempt to

classify the types of sugars present.

Total nitrogen was determined by the Koch and

McMeekin modification of the micro-Kjeldahl method (Hawk

et al., 195^0 using ammonium sulfate as the standard.

Attempts to assay protein utilized the Biuret and the Folin

and Ciocalteau methods with Versatol A as the standard.

Amino nitrogen was measured by the ninhydrin methods (Kabat

and Mayer, 1961) for primary amino acids and for proteins.

Glycine was used to prepare amino nitrogen standards for the

ninhydrin tests.

Spectrophotometrie reading in the ultra-violet light

range were made on the Beckman model DU spectrophotometer

while visible light readings were made using both the Beckman

models B and DU spectrophotometers.

Elemental analysis of the glycopeptlde was performed

by the Galbreath Microanalytical Laboratory.

Hydrolysis

Hydrolysis of the glycopeptlde was accomplished by

dissolving enough glycopeptlde in four volumes of water to

yield, when one volume of 10 N hydrochloric acid was added, the final concentration of 10 mg glycopeptide/ml of 2 N hydro­

chloric acid. For instance, 50 mg glycopeptlde dissolved in

4 ml of distilled water plus 1 ml of 10 N hydrochloric acid

resulted in the desired concentration. The solution was

heated under reflux conditions by either hot plate or boiling

water bath at 90-100 C. The use of a three-necked round-

bottomed reaction flask permitted a thermometer to be in­

serted in one neck, samples to be removed from a second neck,

and attachment of the water-cooled reflux apparatus at the

third neck. Samples taken at various time periods were

evaporated to dryness in an evacuated dessicator over calcium

sulfate, sodium hydroxide, or phosphorus pentoxide in order

to remove the acid. Five to ten cycles of hydration and

evaporation were sufficient to assure that further study of

the glycopeptlde could be performed without interference

from residual acid.

The course of the hydrolytic breakdown was followed

by the colorimetric determination of reducing sugars using

the method of Somogyl (1945)-

Paper chromatography

A variety of solvent systems was used in developing

the one-dimensional chromatograms. Butanol: acetic acid: water « 60:15:25 as prepared by Smith (1958) was most con­

sistently employed for the separation of both amino acids and sugars but the following systems were found useful 21 (all numbers indicate volumes unless otherwise noted):

For the separation of sugars

1# iso-propanol:water = 4:1

2. ethyl acetate:pyridine:water = 12:5:4

3. pyridine:iso-propanol:water:acetic acid =

8:8:4:1

For the separation of amino acids

1. the systems labeled 1, 2, and 3 above

2. phenol:water = 160 g:40.

Whatman #1 chromatography paper (18 1/4 flX 22 1/2") cut

to desired widths was used. The paper was delineated,

spotted with samples and folded according to the protocols

described by Smith (1958).

Chromatograms were developed in Reco chromatocabs or

round glass tanks (24" X 12" diameter) for periods of from

12 to 24 hours. The descending method was always used.

Location of spots was most often accomplished by the

silver nitrate-sodium hydroxide method (Trevelyan et al.,

1940) for sugars and 0.2 per cent w/v ninhydrin in acetone

for substances containing primary amino groups. In addition,

sugar spots were sometimes located by the benzidine, hex-

osamine, sublimating iodine, and para-anisidine reagents.

The first two reagents are described by Smith (1958) and the

latter two may be found in Block et al. (1958). Dip techniques were used in preference to spray methods for spot location. 22

Data were recorded as Rf values when solvent fronts were retained on the paper or as Rg (g » glycine) when

"durchlauf" techniques were necessary.

Column chromatography

Continuous flow column chromatography was employed to a limited degree to separate components of the acid hydrolyzed glycopeptlde. Pyrex columns of 3/4" diameter and 12" long were packed by settling with N,N-Diethylaminoethylcellulose

(DEAE cellulose) which had been prewashed and converted to the borate form (Kundig et al., 1961). The packed column was washed with water until the eluates gave a negative anthrone test. The column was then attached to a Rinco frac­ tion collector set to collect 5 ml fractions of eluate.

Elution was accomplished by introducing Into the column water, increasing molarities of sodium borate ranging from

0.01 to 0.25 M, and finally 0.5 N sodium hydroxide. Each sample was assayed for amino acid material by the ninhydrin test (Kabat and Mayer, 1961) and for the determination of reducing sugars by a submicro method (Park and Johnson,

1949)« When the Park and Johnson test no longer was usable at the higher molarities of sodium borate, the Somogyi (1945) test was used. Samples showing a peak representing amino acid or sugar content were pooled and pervaporated in Visking tubing in front of a draft created by an electric fan. The resulting concentrates were then spotted on paper chromat­ ograms for further identification of glycopeptlde components. EXPERIMENTAL RESULTS

I General properties of the glycopeptlde

Elemental analysis Two different preparations of the pseudomonas slime material were submitted to the Galbreath Microanalytical

Laboratory for elemental analysis. Results of the analyses are shown in Table 2. The results show that despite vari­ ation in composition between the two preparations there existed a constant ratio of eight carbon atoms to each nitrogen atom in both preparations.

TABLE 2

ELEMENTAL ANALYSES OF THE PSEUDOMONAS GLYCOPEPTIDE PREPARATIONS

Preparation #H #N #Fe $Ash C:N Ratio

1 24.42 5.49 3.69 5 .1 6 - 3 6 .0 8:1.015

2 37-61 5.9 5 .8 8 1.3 2 3*09 - 8 :1.0 6

Solubility and pH

The glycopeptlde was soluble in water but not in ethanol, methanol or acetone. It was soluble in solutions having either a basic or acidic pH. The dried glycopeptlde

24 25 went into solution only with difficulty. Each particle of

the dried glycopeptlde Imbibed water and formed a large puff

ball which after agitation slowly went into solution. The

solvation process could be accelerated by placing the

glycopeptide-water mixture in a water bath at 56 C until

solvation was complete. The solution formed was extremely

viscous. A solution containing only 500 pg glycopeptide/ml water still demonstrated a viscosity of 2 .6 7 relative to

distilled water at the same temperature in a Cannon-Fenske L rViscosity of test solution in seconds viscometer tube. [------“— ------= Viscosity of D*H20 in seconds Relative Viscosity].

The pH of the glycopeptlde was determined by dialyz-

ing a glycopeptlde solution (1 mg glycopeptide/ml water)

against 0.0143 N hydrochloric acid followed by thorough dialysis against deionized water of pH 5.4. The glycopep-

tide demonstrated a final pH of 4.8 under these conditions.

Inert components

Samples of the glycopeptlde which had been dried and

stored in a dessicator over calcium sulfate were weighed and placed in tared crucibles in an oven at 105 C. When

the samples attained constant weight, the water lost was

10.51 per cent of the weight. The dry material was then

ignited. After ignition to constant weight, the ash remain­

ing was 7*94 per cent of the weight of the sample. Inert

components thus represent 18.45 per cent of the total sub- 26 stance of the glycopeptlde. Desslcator-stored glycopeptlde was used as the starting material for all tests.

Hydrolytic breakdown

Figure 4 shows a typical curve representing the release of reducing sugars upon hydrolysis of the glycopep- tide with 2 N hydrochloric acid for various time periods.

The appearance of reducing sugars was rapid, reaching high levels at the earliest sampling time. Reducing sugar con­ centrations attained their peak value after only 30 minutes at 100 C and remained at a fairly high level until after four hours at 100 C when the concentration fell rapidly.

At the maximum point reducing sugar represented 73.3 per cent of the material present.

Paper chromatography semi-quantitatively confirmed the maximum appearance of a hexose spot in 30 minute hydrolyzates. The chromatograms further indicated the dis­ appearance of the hexose spot between two and twelve hours.

Amino acid spots appeared on chromatograms of hydrolyzates beginning with the earliest hydrolyzate sample at 10 minutes. The ninhydrin reactive spots were faint in devel­ oped chromatograms of 10 minute hydrolyzates and became more intense in chromatograms of samples hydrolyzed for longer periods of time. 27

25 ■

Time (Hours)

Figure 4.— Presence of reducing sugars during acid hydrolysis of the Pseudomonas Glycopeptlde. 28 II Determination of nitrogen containing components of the glycopeptlde

Total nitrogen

Micro-Kjeldahl determinations of the total nitrogen content of the glycopeptlde Indicated the presence of 3.2^ pg nitrogen/100 pg glycopeptlde. Nitrogen content was calcu­ lated by comparing the optical density reading of test samples against a standard curve of (NHzj)2S02|. Versatol A was used as a control to insure that digestion was complete by the techniques employed.

Biuret test

Biuret tests for protein determination using a

Versatol A standard were negative. This finding confirmed the earlier report by Hrubant and Randles (1958) of a simi­ lar negative Biuret test. However, as subsequent data show, the absence of a positive Biuret test must be viewed in the light of the test's requirement for a pentapeptide (Mehl et al., 19^9) rather than an indication that amino acid con­ taining substances are not present.

Folin and Ciocalteau test

Although Biuret tests were negative it was believed that another means of measuring protein might yield positive results. The Folin and Ciocalteau method was attempted.

The glycopeptlde showed optical density readings so minute 29 (Table 3) when tested by this method that the slime material was considered to be lacking In the cyclic amino acids essential to a true positive reaction In this test. Confi­ dence In the fact that tyrosine was not present In the glycopeptlde was increased by the failure of the glycopep­ tlde to absorb ultra-violet light, a characteristic of peptides and proteins which possess cyclic amino acids. It

Is proposed that the observed optical density readings in the Folin and Ciocalteau test is the result of interferance by reducing sugars present in the glycopeptlde.

TABLE 3

FOLIN AND CIOCALTEAU TEST OF THE GLYCOPEPTIDE

Ug protein Material N Hg N X 6.25 factor % protein 0.5 mg glycopeptlde 2.5 15.6 3.12 1.0 mg glycopeptlde 5.0 31.2 3.12 Average 3.12

Ninhydrin tests

Two tests for amino acid nitrogen using the chromogen formed by 1,2,3 triketohydrindene (ninhydrin) and primary amino groups were applied in an attempt to quantitate the amount of amino acid and amino sugar nitrogen present in a two hour hydrolyzate of the glycopeptlde. The first test, using potassium cyanide to activate a solution of ninhydrin 30 In methyl cellosolve plus acetate buffer (Kabat and Mayer,

1961) resulted In a finding of O .972 pg nitrogen/100 pg hydrolyzed glycopeptlde. Since the above test measured free amino acids and it was highly probable that short peptides

(2-5 amino acids long)vere present, a second ninhydrin test

(Kabat and Mayer, 1961) designed to measure amino nitrogen in protein was used. The second test Indicated the presence of 1.72 |ig nitrogen/100 pg hydrolyzed glycopeptlde. The latter figure may still be conservative since it has been noted that under some conditions only peptides with N-term­ inal glycine will react in this test. A summary of informa­ tion derived from the ninhydrin tests Is presented in Table 4.

TABLE 4

RESULTS OF NINHYDRIN TESTS FOR AMINO NITROGEN pg 2 hour glycopeptlde pg amino N/100 pg hydrolyzate hydrolyzate TJest "Xa “ - Tesf .“

10 1.93

25 O .96 1 .5 6

50 0.95 1.62

100 0.92 1.72 Average 0.95 1 .7 2 aNinhydrin in methyl cellosolve plus acetate buffer activated by potassium cyanide. ^Ninhydrin plus stannous chloride in methyl cellosolve and citrate buffer. 31 Tests for hexosamines. N-acetyl hexosamines and N-acefcyi neuraminic acid------

A test for hexosamines (Dische and Borenfreund, 1950) showed that only 0 .5 per cent hexosamine as glucosamine was present. Despite the small value obtained, the result Is considered valid because (1) the controls employed ruled out interferance from non-deaminated samples and (2) the inclu­ sion of a sample control containing glycopeptlde but no indole eliminated the turbidity of the glycopeptlde solution from consideration. The Blix modification of the Elson and

Morgan test for hexosamines (Gardell, 1958) was negative.

Tests for N-acetyl hexosamines (Reissig et al., 1955) were consistently negative when carried out on various prepara­ tions of the glycopeptlde. Hexosamine and N-acetyl hexosamine determinations were performed on glycopeptlde samples which had been hydrolyzed for times ranging from 10 minutes to

12 hours.

Tests for N-acetyl neuraminic acid were performed using acid hydrolyzed glycopeptlde, intact glycopeptlde, or neuraminidase-treated glycopeptlde. The direct Ehrlich method (Gottschalk, i960) was used on intact glycopeptlde and neutralized acid hydrolyzates while the Aminoff (196I) method was used on neutralized acid hydrolyzates and neuraminidase-treated glycopeptlde. In no case were sialic acids detectable. 32 III Determination of Carbohydrate Componenta of the Glycopeptlde

Ultra-violet absorption test

A spectrophotometric method (Ikawa and Niemann, 19*19) was used to Identify Individual sugars. In Figure 5 absorp­ tion spectra are presented representing the ultra-violet absorption of the glycopeptlde and various carbohydrates after treatment with 33 per cent sulfuric acid. It may be seen that the peak of absorption at 295 mp. shown by the glycopeptlde more nearly resembled the peak of galactose than of any of the other carbohydrates used as standards.

On the other hand, the glycopeptlde curve between 240 and

260 mp. was most like that of the glucose absorption curve.

Earlier chemical tests had indicated the presence of galactose but after completion of this test it was sus­ pected that both glucose and galactose might be present.

Cysteine test

Data were obtained in the cysteine test (Dische,

1955) which further emphasized the possibility that both glucose and galactose were present in the glycopeptlde.

Figure 6 shows again that in one part of the spectrum (375-

425 mp.) the absorption curve of the glycopeptlde more closely resembled the curve of the galactose while in another part of the spectrum (550-625 mp.) glucose was suggested. 33

0.30

0.25

0.20 Optical Density

0.15

0.10

220 260 300 320 360360 Wavelength ayi

. Glycopeptlde; ® Glucose; a Galactose.

Figure 5 . Ultra-violet absorption spectra of the glycopeptlde and known sugars in the Ikawa and Niemann (1949) test. a.00 roo too

. Glycopeptlde; © Galactose; a Glucose

Figure 6.— Absorption spectra of the glycopeptlde and known sugars In the Dische cysteine test (1955;. 35 Classifying cysteine test

Another cysteine test (Dische, 1949) was used to determine the classes of sugars present in the glycopeptlde.

In Figure 7 the absorption curve of the glycopeptlde is compared to that of galactose, fucose, and sedopheptulose.

The glycopeptlde curve closely resembled the hexose curve as represented by galactose and shows no similarity to either the 6-deoxyhexose or heptulose curves or to pentose curves.

The latter group showed an extremely high peak of 390 mp. for arabinose and at 400 mp. for xylose. Additional information may be garnered from this test by the evaluation of the optical density at 390 mp. minus optical density at 421 mp. for the various sugars. The value 0D 390-421 mp. is shown in Table 5 for representatives of the aldopentose,

6-deoxyhexose, aldohexose and heptulose classes of sugars.

It may be seen that the choice of these particular wave­ lengths yielded a zero, small negative, or positive reading for aldohexoses. Thus other sugars may be determined des­ pite their presence In polysaccharides containing high con­ centrations of aldohexose. The data in Table 5 led the author to eliminate the possibility of the presence of pentoses in the glycopeptlde and to consider doubtful the presence of

6-deoxyhexoses or heptulose. In addition, interpretation of the curves In Figure 7 strongly suggested that neither

6-deoxyhexose or sedoheptulose was present in the glycopeptlde. 36

* i

.0-

. Glycopeptlde; & Galactose; ® Fucose; a Sedoheptulose

Figure 7«— Absorption Spectra of the glycopeptlde and known sugars In the Dische classifying cysteine test (1955). 37> TABLE 5

IDENTIFICATION OF CLASSES OF SUGARS BY THE VALUE OD 390 OD 421 mp. IN THE DISCHE CYSTEINE TEST (1955)

Optical Density at mu Sugar 390 421 390-421 fucose 0.159 0.113 0.046 rhamnose 0.055 0.029 0.026 arablnose 1.65 0.2 2 8 1.422 xylose 1.6 0.204 1.396 glucose 0.273 0.273 0.0 mannose 0.685 0.68 0.005 galactose 0.392 0.409 -0.017 sedoheptulose 0.221 0.249 -0.028

glycopeptlde 0.331 0.35 -0.019

A test similar to the cysteine test but using carba- zole to form the chromogen (Dische, 1955) showed that among

the three major hexoses present in naturally occurring materials— glucose, galactose, and mannose— the absorption

spectra of both glucose and galactose were similar to that

of the glycopeptlde while mannose exhibited an entirely

different curve (Figure 8). The ratios of optical density

535/440 for the three aldohexoses and the glycopeptlde are

shown in Table 6. According to Dische the difference in

ratios of the three aldohexoses permits unknowns to be

identified by comparison to standards. It may be seen in

Table 6 that the OD 535/440 for the glycopeptlde falls mid­

way between that of glucose and galactose in one test. In

a second trial the ratio of OD 535/440 for the glycopeptlde

again fell between that of glucose and galactose but closer

to the former. 38

O.JO ■.

0.0^ ■.

600

. Glycopeptlde; <$ Glucose; & Galactose; q Mannose

Figure 8.— Absorption spectra of the Glycopeptlde and known sugars In the Dische carbazole test (1955). 39 TABLE 6

IDENTIFICATION OF SPECIFIC ALDOHEXOSES BY THE RATIO OD 535/440 IN THE DISCHE CARBAZOLE TEST (1955)

OD 535/440 Test Substance Trial I Trial II glucose 2.35 2.5 galactose 1.25 0.89 mannose 0.71 0.66 glycopeptlde 1.9 2.2

Quantitation of hexoses

Early trials of colorimetric hexose determinations had been performed using glucose as the standard. It was now believed advisable to repeat these experiments using both glucose and galactose as the standards.

Four separate tests designed to quantitate hexoses were applied using standard curves of both glucose and galactose. Results of three of the tests are shown in

Table 7. The results as shown are incomprehensible if they are evaluated on the premise that only one hexose is present. However, if it is assumed that glucose and galactose are present in equivalent proportions, then the calculations shown in Table 8 demonstrate a close agreement in a value of 56.1 per cent total hexose present in the glycopeptlde. The fourth test, based on cysteine (Dische,

1955)> gave consistently lower results. When calculated in 40 the manner of Table 8 the cysteine test showed a final

hexose content of 43*8 per cent.

TABLE 7

QUANTITATION OF HEXOSES PRESENT IN THE GLYCOPEPTIDE BY MEANS OF COLORIMETRIC TESTS

# Hexose in the Glycopeptlde Expressed as: Test Glucose Galactose anthrone 45-5 7 0 .0

Indole 68.5 45.6

tryptophane 53-5 53-5

TABLE 8

CALCULATION OF HEXOSES PRESENT IN THE GLYCOPEPTIDE AS 50:50 GLUCOSE:GALACTOSE

# Hexose in the Glycopeptlde Calculated as: Test 50# Glucose + 50# Galactose = Total Hexose anthrone 22.8 35 57.8

indole 34.25 22.8 57.1 tryptophane 2 6 .8 2 6 .8 53.6

Average 5 6 .1

Glucostat determination Both chromatographic and colorimetric evidence had

indicated that glucose and galactose were present in

approximately equal amounts. In order to determine more

exactly the amounts of each hexose present, a glucostat 41

determination was used. Glucostat (Worthington Biochemical

Corporation, Freehold, New Jersey) is a kit which contains

glucose oxidase and a chromogen. A 30 minute hydrolyzate

of the glycopeptlde was tested by the Somogyi (1945) reduc­

ing sugar method and found to contain 19*1 Mg of reducing

sugar as glucose/100 pg glycopeptlde. An equivalent amount

of the hydrolyzate was then tested by the Glucostat method

and found to contain 9*33 Mg glucose. Glucose and Galaotdsfe

were used as standards. Table 9 shows the final values

obtained for glucose and galactose as a result of the appli­

cation of the Glucostat method. The figures represent 50.9

per cent of the reducing sugar as glucose. By taking the

difference, galactose must then represent 49.1 per cent of

the hexose.

TABLE 9 GLUCOSTAT DETERMINATION OF GLUCOSE PRESENT IN THE GLYCOPEPTIDE

Somogyi Test Glucostat Mg reducing sugas as glu- Mg glucose/ OD 540 mM> cose 100/Mg OD 410 100 Mg gly- Test Substance glycopeptlde mu copeptide 25 Mg 0.31 100 pg glucose O.I85

100 Mg hydrolyzate 0.073 18.5 0.125 10.1

200 pg hydrolyzate 0.1 5 8 2 0 .0 0.237 9.56

50 Mg galactose 0.163 0 .0 2 42 Text for hexuronic acids

Since the glycopeptide demonstrated an acidic pH

reaction it was expected that hexuronic acids might be

present. The Dische carbazole tests (1947a, b, 1950) were

employed to determine the presence or absence of the

hexuronic acids. In Figure 9 is shown an absorption curve

for glucuronic acid and for glucose. The galacturonic acid

and poly-galacturonic acid curves were almost identical to

that of glucuronic acid. Alginic acid (poly-mannuronic

acid) and chondroitin sulfate reacted to a much lesser degree

but showed the characteristic uronic acid peak around

535-540 mu. Substances with an acid function on carbon

number one (glucosamlnic acid, gluconic acid) and on both

carbons numbers one and six (potassium hydrogen saccharate)

did not react in this test. The curve shown for glucose is representative of those for galactose and mannose. It is

obvious from Figure 9 that the glycopeptide showed no uronic

acid peak but instead closely resembled the absorption curve

of aldohexoses.

IV Paper chromatography of acid hydrolyzed glycopeptide

Early attempts to separate sugar components on paper

chromatograms resulted in consistent observation of a hexose

spot plus several other small Rp spots. The hexose spot

repeatedly appeared between glucose and galactose or closer

to one or the other of the two hexoses (Figure 10)• However, 43

0 .3 0 Ootictl 'entity

0.20

. Glycopeptidej 0 Glucose; & Glucuronic Acid

Figure 9. Absorption spectra of the glycopeptide, glucose and glucuronic acid in the Dlsche carbazole test for heuronlc acids (1955)* 44

Spots represent left to right: glucose, galactose, hydrolyzate of glycopeptide, hydrolyzate of earlier prep­ aration of glycopeptide, N-acetyl neuraminic acid.

Solvent system: butanol:acetic acid:water = 60:15:25.

Method: descending in Reco Chromatocab 17 hours. Spot Location: silver nltrate-sodium hydroxide method of Trevelyan et al. (1950).

Figure 10.— Paper chromatogram of glycopeptide hydrolyzates and known carbohydrates. 45 when a suitable solvent system was eventually found, the

hexose spot was resolved into two separate spots represent­

ing glucose and galactose. A solvent system composed of

iso-propanol: water =4:1 gave partial separation of the two

hexoses (Figure 11). Finally a solvent system made up of

ethyl acetate:pyridine:water = 12:5:4 completely resolved

the hexose spot into two spots equivalent to glucose and

galactose (Figure 12). Visually the two spots appeared to

represent equal quantities of each hexose.

It is of further interest that two spots appeared

regularly in addition to the hexose spot. Both migrated

shorter distances than the hexose spot and were much fainter

than the hexose spot. Note In this respect the absence of

these two small Rp spots in Figures 11 and 12 where smaller

quantities of the hydrolyzate were placed at the origin.

The spot appearing just short of the hexose spot in Figure

10 persisted In chromatograms of hydrolyzates of long dur­ ation (12 to 24 hours In 2 N hydrochloric acid) while the hexose spots had disappeared. In Rp value and In resistance

to hydrolysis spot 2 resembled an amino sugar but never matched the known amino sugars run as standards (Table io).

The hexosamine most nearly approximated by spot 2 was glucosamine hydrochloride as shown in Table 11. A charac­ teristic of spot 2 was a frequent slowness in developing a dark spot in response to the silver nitrate-sodium hydroxide Spots represent left to right: glucose, hydrolyzate of glycopeptide, galactose.

Solvent system: iso-propanol:water = 4:1.

Method: Descending in round glass chromatography tank 21 hours.

Spot location: Silver nitrate-sodium hydroxide method of Trevelyan et al. (1950).

Figure 11.— Partial separation of the glucose and galactose components of the glycopeptide on paper chromatograms. 47

Spots represent left to right: glucose, hydrolyzate of glycopeptide, galactose.

Solvent system: ethyl acetate:pyridine-.water = 12:5:4.

Method: descending in round chromatography tank 21 hours durchlauf technique.

Spot location: silver nitrate-sodium hydroxide method of Trevelyan et al. (1950).

Figure 12.— Complete separation of the glucose and galactose components of the glycopeptide on paper chromatograms. TABLE 10 48 COMPARISON OP THE Rp VALUES OP HEXOSAMINES TO THE Rp VALUE OP A SPOT APPEARING IN CHROMATOGRAMS OP HYDROLYZED GLYCOPEPTIDE

Test Substance Rglucosamlne glucosamine 1.00 galactosamlne 0.93 N-acetyl glucosamine 0.88 Unknown spot 1.8

Solvent System: butanol:ethanol:water = 4:4:1.

Method: descending In Reco Chromatocab 19.5 hours.

Spot Location: sliver nitrate-sodium hydroxide method of Trevelyan et al. (1050).

spot location technique. The shortest running spot in the hydrolyzate shown on the left in Figure 10 (Rf = 0.15). The

Rf = 0.13 spot appeared only in chromatograms of short dur­

ation (to 30 minutes) hydrolyzates. A spot with these

characteristics could easily represent a disaccharide not

yet broken down by the hydrolytic process. However, in

other solvent systems this small Rf spot ran too far to be

considered a disaccharide (Table 11).

In Table 11 are presented Rf values for substances vhich reacted with a silver nitrate-sodium hydroxide reagent

(Trevelyan et al., 1950). The most easily recognized spot was *the "hexose" spot called spot 3» As reported above and shown in Figures 11 and 12 the hexose spot could be resolved

into two spots representing glucose and galactose. Spot 2 most nearly represents glucosamine hydrochloride. Spot 1

could not be further identified. TABLE 11

COMPARISON OF Rp VALUES OF SILVER NITRATE-REACTIVE SPOTS ON PAPER CHROMATOGRAMS USING VARIOUS SOLVENT SYSTEMS

Solvent System Substance 1 2 3 glucose 0 .5 6 0.24 0.42 galactose 0.23 0.41 mannose 0.47 glucosamine *HC1 0.49 0 .2 0

N-acetyl glucosamine 0 .6 6 0.37 Glycopeptide hydrolyzate

Spot 1 0.46 0 .1 1 0 .2 8

Spot 2 O .50 0 .1 8 0.33

Spot 3 0.57 0.23 0.43 Solvent Systems: 1. Pyridine:lso-propanol:water:acetic acid = 8:8:4:1.

2. Butanol:acetic acid:water - 60:15:25*

3* Iso-propanol:water = 4:1.

Method: descending in Reco Chromatocab or round glass tanks for 12 to 24 hours.

Table 12 lists Rf values for an array of ninhydrin reactive substances which were used as standards in an attempt to identify amino acids or amino sugars in hydroly- zates of the glycopeptide. Consistent confirmation of the presence of glycine and alanine may be found regardless of the solvent system used. Other ninhydrin reactive spots 50

TABLE 12

COMPARISON OF RF VALUES OF NINHYDRIN REACTIVE SPOTS ON PAPER CHROMATOGRAMS USING VARIOUS SOLVENT SYSTEMS

Test Substance 1 2 3a 4 5 6 glycine 0.24 0.25 1.0 0.29 0.11 0 .2 6 alanine 0.37 0.31 1*5 0.4 0.17 0.42 arginine*HC1 0 .1 9 0.15 0.10 aspartic acid 0.15 0.23 0 .7 0 asparagine 0.16 0.67 0.2 glutamic acid 0.22 1.17 glucosamine *HC1 0.49 1.92 0.45 lysine 0.18 0.12 0.06 arginine free base 0 .1 9 0.19 amino tyrosine*HC1 0.35 tryptophane 0.5 leucine 0 .6 1 0.70 methionine 0.53 tyrosine 0.54 phenylanlanine 0.57 proline 0.39 hydroxypro1ine 0.34 valine 0.51 isoleuclne 0.68 51 TABLE 12— continued

Vest Substance 1 2 3a 4 5 6 glycopeptide hydrolyzate

Spot 1 0.14 0.15 O .6 5 0.13

Spot 2 0.18 0.17

Spot 3 0.24 0.25 1.04 0.31 0.11 0.2 6

Spot 4 0.36 0.31 1.47 0.42 0.17 0.40

Spot 5 0.49/0.54b 0.6 I .8 9 0.59 0.3 Spot 6 2.2

Spot 7 0 .6 1 0.75 2 .6 2 0.71 0.4 0.74

aRglycine values since front was not retained on paper. b().491 in 30 minute hydrolyzates and 0.54 in 24 hours hydrolyzates.

Solvent Systems: 1. Pyridine:iso-propanol:watersacetic acid =» 8:8:4:1. 2. Butanol:acetic acid:water = 60:15:25.

3. Same as solvent system 2 but durchlauf technique used.

4. Iso-propanol:water = 4:1.

5. Ethyl acetate:pyridine:water = 12:5:4.

6. Phenol:water = 160 g:40.

Method: descending in Reco Chromatocab or round glass chromatographic tanks for 12 to 24 hours.

Spot Location: dip In 0.2 per cent ninhydrin followed by heating at 100 C for 5 to 7 minutes. appeared in various trials depending on the amount of sample

placed at the origin, the solvent system or the duration of

run. For instance, solvent system 2 in Table 12 showed a

spot resolved from the spot labeled spot 1. Again in sol« vent system 3 where a long run was obtained by allowing the

solvent front to drip off the end of the paper (durchlauf

technique) a new spot appeared. From visual observation of

the spots appearing in the chromatograms it may be stated

that glycine and alanine are the major components of the peptide moiety of the slime material and that there are in addition three other minor constituents. The minor con­ stituents most nearly resemble lysine or arginine in the case of spot 1 and spot 2 and leucine for spot J. Glutamic acid was also suggested. Spot 5 does not consistently follow the pattern of glucosamine hydrochloride but does suggest it in solvent systems 1 and 3.

V Column chromatography

Column chromatographic studies were carried out employing the techniques described under Materials and

Methods. Results were generally unsatisfactory. Separation occurred to some extent as demonstrated by the appearance of peaks representing the localization of reducing sugars and primary amino group compounds into certain samples or groups of samples. However, when the samples composing the peaks were concentrated and examined by paper chromatographic means, no unique fractions were evident. Failure to obtain good results was probably attributable to the application of more sample to the DEAE cellulose column than it was capable of firmly adsorbing and selectively eluting. Figure 13 shows the peaks obtained when graphs were prepared of optical density reading of reducing sugar and ninhydrin tests against the volume of eluate introduced into the column. Paper chromatography showed that each successive peak contained a hexose spot plus the two major ninhyrin spots— glycine and alanine. 54

i.e

0.6

pttesl Witty 0.U

. __ ■---- 1-- ^-1-h f »«t«n ».o»iV osm 0.075 o.m i O.OJS O.S 0.5N K.OM

. Reducing sugars (Park and Johnson, 1949).

Ninhydrin test (Kabat and Mayer, 196l).

Column 3/4" by 12" packed with N, N-Diethylaminoethyl cellulose and converted to the borate form (Kundlg et al., 1961).

Figure 13*— Determination of reducing sugars and ninhydrin reactive material in samples of eluate from a DEAE Cellulose column. DISCUSSION

Nomenclature

In this dissertation it was considered that the best

though still imperfect name to be applied to the purified

slime material was the term glycopeptide. A great deal of

confusion surrounds the question of the proper nomenclature

for carbohydrate-protein complexes and hexosamine-containing

substances. Winzler (1958) attempted to end some of the

controversy by proposing terminology for hexosamine contain­

ing substances. In his nomenclatural system, Winzler estab­

lished a group called glycoproteins which was to consist of

those hexosamine-containing substances which lacked hexuronic acids and had relatively small amounts (0.5 per cent or more) of hexosamine firmly bound to protein and in which the substances as a whole displayed the properties of proteins.

The pseudomonas slime material discussed in this dis­ sertation fits Winzler's category of glycoproteins in that it lacked hexuronic acid and contained only small amounts of hexosamine. On the other hand the negative protein tests and the positive peptide tests demanded that the suffix peptide rather than protein be used. Furthermore, the char­ acteristics of the glycopeptide, especially in regard to

55 56 ease of hydrolysis, are closer to those of carbohydrate

than to protein.

Morphological localization of the glycopeptlfe

The glycopeptide material has been asserted to be a

component of the slime-coating surrounding the organism.

The assertion was based upon the fact that the glycopeptide

is easily removed from the cell. Preliminary chemical

studies, however, indicated that the material isolated from

the pseudomonas was a complex substance which in many ways

resembled cell wall material more than the generally simpler

capsular material of microorganisms. Salton (i9 6 0), for

instance, has demonstrated that the Gram-positive organism

Micrococcus lysodeikticus has a cell wall composed of hexosamine, glucose, muramic acid and the four amino acids alanine, glutamic acid, glycine, and lysine, and that the cell walls of Gram-negative organisms consist of a protein- polysaccharide-lipid complex and a mucocomplex.

Attempts to visualize a capsular or slime layer sur­ rounding the Pseudomonas organisms met with mixed results.

Using (1) India ink negative staining methods or (2) adjust­ ment of the pH of the suspending fluid followed by dilute basic dyes, results could easily be interpreted either to mean that only a small amount of capsule was present or that it was entirely absent. Despite the fact that attempts to visualize a capsule were not conclusive* several lines of

reasoning point to an extracellular location of the glyco­

peptide. Chief among these is the ease with which the

material is isolated from the cells. In essence it is only

necessary to beat the slime layer off the bacterium and

precipitate the slime layer material from its water solution.

If the glycopeptide were derived from the cell wall or the

interior of the cell, much more severe means would be needed

to disrupt the integrity of the cell than those used in the

course of this research. Furthermore, if the material were

from the cell soma, preparations might be expected to contain

protein, uronic acids, pentoses and other somatic constitu­

ents . None of the latter substances were detected in the

final purified glycopeptide preparations. Thus the arguments

for an extracellular localization of the glycopeptide appear

to effectively outweigh those for a somatic localization.

Techniques employed

The two basic techniques employed in this study were

colorimetric testing and chromatographic separation. At various times it was felt that results from one or the other of the techniques were more reliable. The respective value of each technique is worth discussing both with respect to the results obtained and as a guide to future work on similar problems. Much of the data reported under experimental results

were derived from colorimetric determinations based on the

tests of Dische (1955)* Perusal of the paper by Dlsche

leads one to believe that the composition of a polysaccha­

ride may be determined in detail by the use of one or a

combination of colorimetric tests. Otoe degree of validity

in this approach is sharply pointed up in the results

obtained in the course of this study. Seldom did an individ­

ual colorimetric test give so clear-cut a result as to

remove doubt that further testing was necessary. An example may be found in Figures 5-7 where only the minutest differ­

ences are found in the shape of the absorption curves and

Imagination must be compounded with experience to find that

both glucose and galactose are present rather than one or

the other of the hexoses. On the other hand Figure 8

clearly rules out the presence of hexuronic acids and

Figure 9 permits certain assertions to be made with confi­ dence. Otoe value of colorimetric tests was excellently

exhibited by the eventual confirmation by the Glucostat method of an assumption made by calculating the results of

three separate colorimetric tests. The majority of the

Dische tests have distinct advantages: (1) the test material need not be hydrolyzed as is the case in chromatography,

(2) certain components may be determined in the presence of an excess of other substances, (3) quantitation was achieved 59 In addition to qualitation, and (4) the procedures are more

convenient than paper chromatography.

Proponents of chromatographic methods for the eluci­

dation of composition claim a greater degree of exactness

for their favored technique. Yet an identical value is

merely an indication that two compounds have migrated to the

same degree and is not a proof of Identity. Still, colori­

metric tests alone would not have so clearly provided the

information in Figure 12 that glucose and galactose were

both present.

Ultimately it must be accepted that both techniques

are of value in a study of chemical composition. It is only

hindsight which permits one to say that one technique should have been used more or less than another or that the results

of one type of test is to be more weighted than results from the other.

Chemical composition of the pseudomonas glycopeptide

A summary of the findings described under experimental results is presented In Table 13. The value for carbo­ hydrates is taken from Table 6. Amino acids were calcu­ lated on the premise that all the nitrogen measured by the mlcro-Kjeldahl method was amino acid nitrogen since the nitrogen in the trace amount of amino sugar was negligible. 6o

TABLE 13

SUMMARY OP DATA ON THE CHEMISTRY OP THE PSEUDOMONAS GLYCOPEPTIDE

Constituent % observed® Carbohydrates 56.1

Glucose 50.9 # of the hexose present

Galactose 49*1 # of the hexose present

Amino Acids (3*24 # nitrogen X 6.25 conversion factor ) 20.2

glycine Principal amino acids present alanine

leucine

lysine possibly present in small amounts arginine

glutamic acid

Amino sugar as glucosamine 0.5

Inert components 18.45

Water 10.51 %

Ash 7*94 56

Total 95.25 aPigures in this column are percentage by weight of the dessicator-dried glycopeptide preparation. 61

It has been shown that glucose, galactose, glycine and alanine are the major constituents of the glycopeptide.

Elemental analyses have Indicated, in addition, a ratio of eight carbon atoms to each atom of nitrogen. Collating the above facts, we may postulate a structure which would con­ tain the repeating unit represented as follows:

glucose - galactose / glycine

alanine /

The postulated repeating unit has the formula

^17^2 8^2^12 and a m°lecu^ar weight of 452 of which 7 1 .7 per cent of the weight is contributed by glucose and galactose and 2 7 .3 per cent by glycine and alanine. Since inert com­ ponents have been shown to represent 18.45 per cent of the glycopeptide, the values of 7 1 .7 per cent for hexose and

2 7 .3 per cent for amino acids must be corrected to 5 8 .6 per cent and 22.3 per cent for hexoses and amino acids respec­ tively in order to properly value the contribution of the inert material in the glycopeptide. Comparison of the values calculated on the basis of the proposed repeating unit with the values observed in quantitative tests shows good agreement. Furthermore, the postulated structure would have 8 .5 carbon atoms to each atom of nitrogen— close to the observed value of 8:1. Some of the slightly higher 62

carbohydrate and amino acid value found in the calculated

figures may be accounted for by the amino sugar unit which

could possibly be present as a terminal immunologicly

determinative unit (Cf. next section). The presence of this

last mentioned unit with its proposed formula of C^qHi j ^ O q

would help to reduce the ratio of carbon to nitrogen atoms

from its calculated 8.5:1 ratio closer to the 8:1 ratio

found in elemental analyses.

Again referring to the proposed repeating unit, a

comparison may be made of the values determined by elemental

analyses and the figures calculated on the basis of the

postulated repeating unit having the formula 03^ 2 8^2 0 1 2*

It may be seen in Table 14 that the figures calculated on an

assumed C17H28N2O12 formula agree fairly well with the analyses of earlier preparations of the glycopeptide. How­

ever, a marked and difficult to explain discrepancy exists

between the calculated amount of nitrogen and the nitrogen actually measured by micro-Kjeldahl methods. If, for some reason the observed micro-Kjeldahl value was too low, it is easy to accept the fact that the calculated nitrogen value

of 5 -0 6 per cent representing 31*6 per cent amino acids is

valid. Acceptance of this line of reasoning would put the

total arrived at in Table 13 at 106.65 per cent— a not unreasonable figure in work of this type.

The nature of the linkages present, though not directly determined, can be inferred to some degree. The 63 TABLE 14

COMPARISON OP ELEMENTAL ANALYSES WITH CALCULATED ANALYSES FOR A C17H2 8N2O12 UNIT

C17 «28 n2 °1 2 % Weight contributed3 45*0 6 .2 6 .2 42.4

Galbreath Analysis 1 24.42 5.49 3.69

C17H28^2 °1 2 corrected for 36 # ashb 2 8 .8 3*97 3*97

Galbreath Analysis 2 37.61 5.9 5 .8 8

C17H2 8N2 °1 2 corrected for 9 .2 % ashc 40.9 5.64 5*64

c17h2 8n2 °1 2 corrected for 18*45 # inert material3 36.7 5 .0 6 5 .0 6

Calculated as percentage of molecular weight of Cl7H28Nfi0i2 . bValue for ash reported by Galbreath Microanalytical Laboratory for preparation 1. cAsh in preparation 2 estimated in proportion to lower percentage of sulfur in preparation 2 than in preparation 1.

^Observed value for inert material reported under experimental results. amino acids present are most probably short chains of no more

than four amino acids since the Biuret test would have measured chains of five or more amino acids. The amino

acids are probably linked to each other by standard peptide

linkages since two hour hydrolyzates tested by two types of

ninhydrin test indicated that the amino acids were still

present in strongly bonded form. These data agree with the

accepted view of the relative resistance of the peptide bond

to the hydrolytic conditions employed. The manner in which

the peptide chains are bound to the carbohydrate may be in

any one of at least three different types as shown in

Figure 14. Of the three possible linkages shown in Figure

14, form I is quite common in microbial cell wall, blood group substances and other polysaccharide-peptide complexes.

Form II has been found In nature and synthesized in many

variations in the laboratory. A linkage* similar to form III

is found in muramlc acid. A muramic acid type linkage in which the carboxyl group of the amino acid Is not involved

in the bond would help to explain the acidic nature of the glycopeptide. The indicated presence of the dicarboxylic amino acid, glutamic acid, would contribute other free carboxyls to the production of a final acidic reaction.

There Is no question that the peptide is firmly linked to the carbohydrate and not merely loosely associated with it. When the glycopeptide was treated with trypsin and 65

c 0 1 II C 0 H-O-C C t 11 i i C-O-C-C-R H-C-O-C i t i i C N-H H-N C i t i t C H H C i i C C t I c Ester Linked III

Ether Linked

C H 0 1 I It C-N-C--C-R i I c N-H I i c H t c 1 c

II

Peptide Linked

Figure 1*1.— Possible linkages of amino acids to the carbohydrate in the glycopeptide. 66

thenextracted by the Sevag method the entire molecule passed

into the phase containing the peptide rather than the

occurrence of a separation of carbohydrate into the aqueous

phase and the peptide into the ether phase.

The peptide moiety of the glycopeptide shows a remark­

able degree of resistance to proteases due either to the

(1) lack of the specific amino acids comprising the pair in

a peptide bond, (2) presence of essentially one amino acid

per hexose in an ester linkage, or (3) shortness of the

peptide chain.

The carbohydrate linkages are most likely glycosidic

in nature. The 1:1 ratio of glucose to galactose suggests

a lactose unit. The rapidity with which the glycopeptide is

broken down by acid hydrolysis coupled with the fact that

acid hyrolysis is known to more rapidly sever (1-4) linkages

than (1-6) linkages, further suggests that the glucose and

galactose are linked by the (1-4) linkage found in lactose.

Examination of linkages by attempted enzyme degrada­

tion showed that neither pectinase, neuraminidase, lysozyme,

nor hyaluronidase caused significant release of reducing

sugar or reduction of viscosity of the glycopeptide. The

absence of the specific substrates is indicated.

Rh0(D) specificity

Part of the original impetus to determine the struc­

ture of the pseudomonas glycopeptide was the observation by Dodd et al. (i9 6 0) that the glycopeptide inhibited the reaction between Rh0(D) antibodies and Rhc (D) erythrocytes to a degree either matching or surpassing the ability of

N-acetyl neuraminic acid to inhibit the system. It has been shown that the ability of a substance to block a serological reaction often indicates some form of similarity between the specific material and the interfering material. Kabat (1958) emphasized the value of using cross-reactions as an indica­ tion of similarity in linkage or constitution. On the basis of such reasoning it was believed that examination of the chemical structure of the glycopeptide might uncover sialic acid, a similar compound, or a linkage that would not only explain the glycopeptide's ability to inhibit the Rh0(D) system but also give an insight into the structure of the

Rh0(D) determinant group. A review of the pertinent informa­ tion may be found in Dodd et al. (i9 6 0). As described under experimental results, sialic acid was never detected by any of the techniques employed. Experiments were designed to detect the sialic acid whether it was present as a terminal group or as an internal component of the glycopeptide. Thus, tests on neuraminidase treated, 10 minute hydrolyzates, or intact glycopeptide should have been positive for sialic acids had they been present in a terminal position; tests on longer term hydrolyzates and on enzyme degraded glycopeptide should have detected internally situated sialic acid. As reported under experimental results no sialic acid was

detected. Assuming that sialic acid is absent from the

glycopeptide, other explanations for the glycopeptlde1s

ability to inhibit the Rh0(D) system are (1) a structure

similar to sialic acid or (2) involvement of a linkage simi­

lar to that determining Rh0 (D) specificity. A structure which distantly resembles that of sialic acid is shown in

Figure 15 where it is also compared to the structure for the

cell wall material, muramic acid. The structures have in

common the N-acetyl substitution and a carboxyl function.

These similarities might be sufficient to confer cross­

reactivity. Substantiating evidence for the presence of such

a structure is extremely tenuous. The meager evidence con­ sists of the fact that the compound labeled "silver nitrate reactive spot number 2" in Table 11 displayed a resistance

to hot mineral acids in line with that reported for muramic acid. The resistance of muramic acid to acid hydrolysis has been attributed to the presence of a compound having a free carboxyl group substituted by means of an ether linkage to a secondary hydroxyl group of a hexose. Furthermore, the behavior of spot 2 in developing color tardily when treated with silver nitrate might indicate unusual substitution since the silver nitrate reaction depends on reactivity with free hydroxyl groups. 69

0 0 11 it C-OH C-OH i i — C-OH C-H i i H-C-H CH3 i H-C-OH H-C=0 t 1 H-C-NH-C-CH: H-C-NH-C-CH3 i ii 1 ii J — C-H 0 — C-H 0 i t H-C-OH H-C-OH i 1 H-C-OH 0 H-C-OH i 11 t c h 2o h C-OH CH20H 1 — C-H Sialic Acid 1 Muramic Acid H-N-H

H-C=0 1 H-C-NH-C-CHq 1 «iti -j — C-H 0 1 H-C-OH 1 H-C-OH t CH20H

Glycine ether-linked to N-acetyl glycosamine

Figure 15 •— Comparison of Sialic Acid, muramic acid and a proposed glycine-N-acetyl glucosamine compound. In regard to linkage, it has been found that sialic acid in a human milk fraction is attached to lactose (Gyorgy,

1958) while in ox brain mucollpid (Rosenberg et al., 1 9 5 6) sialic acids appear to be attached to galactose. In the glycopeptide there Is a sufficient amount of galactose and a nearly 1:1 ratio of glucoBe to galactose as found in lactose to provide the reportedly essential linkage compounds for a sialic acid-like substance. The entire discussion of serologic specificity must, of course, be viewed as speculative. SUMMARY

The chemistry of a slime material from a species of

Pseudomonas isolated from highly acid water was examined by

colorimetric and chromatographic methods. Preparation of

the slime material for study was achieved by agitation in a

Waring Blendor, removal of cells and alcohol precipitation

of the slime material from water solution.

Upon examination, the slime material was character­

ized as a glycopeptide and found to consist of 5 6 .1 per cent

aldohexoses, 2 0 .2 per cent amino acids. 0 .5 per cent hexos-

amine as glucosamine and 18.^5 per cent inert material. A

combination of colorimetric, enzymatic, and chromatographic

methodology detected and then confirmed that the hexoses

present were glucose and galactose in an almost exact 1 :1 ratio•

Although the glycopeptide did not yield a positive

Biuret or Folin and Ciocalteau test and did not absorb

ultra-violet light, amino acids were detected by ninhydrin

tests and chromatographic techniques. Chromatography showed

the presence of glycine and alanine and indicated that leucine, lysine, arginine and glutamic acid were also present in trace amounts.

71 72 The absence of hexuronic acids, pentoses, 6-deoxyhex-

oses, and N-acetyl hexosamines was indicated by colorimetric

tests and by evaluation of absorption spectra.

Sialic acid was not detected by any of a number of

techniques. It had been proposed that sialic acid might be

present in the glycopeptide since all compounds previously

found to inhibit the Rho(D) human blood-group system con­

tained sialic acid. Since the glycopeptide did inhibit the

Rho(D) system, other explanations for its serological

activity were proposed.

The knowledge that glucose, galactose, glycine and

alanine were the chief components of the glycopeptide led to

the postulation of a repeating unit whose formula and

structure compared closely to the experimental results observed in the course of the study. A discrepancy in nitro­ gen values did exist, however.

A speculative structure which both accords with quanti­

tative data and proposes an explanation for the serological activity is shown in the discussion section. It consists of a repeating unit of glucose linked (1-4) to galactose with chains of one to four amino acids long attached in ester linkage to a secondary hydroxyl of a hexose. Terminally located is an N-acetyl glucosamine with a glycine O-substi- tuted in an ether linkage closely resembling the sialic acid or muramic acid type structure. BIBLIOGRAPHY

Aminoff, D., 1961 Methods for the quantitative estimation of N-acetyl neuraminic acid and their application to hydrolysates of sialomucoids. Biochem. J., 81, 384-392.

Aminoff, D., Morgan, W.T.J., and Watkins, W.M., 1952 Studies in iraraunochemistry. II The action of dilute alkali on the N-acetylhexosamines and the specific blood-group mucoids. Biochem. J., J51, 379-389.

Barry, G.T., and Goebel, W.P., 1957 Colominic acid, a sub­ stance of bacterial origin related to sialic acid. Nature, 179, 206-207.

Bigley, N.J., Dodd, M.C., Randles. C.I., Geyer, V.B., and Lazen, A.G., 1963 The Rh0(D) specificity of a poly­ saccharide isolated from a species of Pseudomonas. J. Immunol., j}0, 526-531.

Block, R.J., Durram, E.L., and Zweig, G., 1958 A Manual of Paper Chromatography and Paper Electrophoresis, Academic Press Inc., New York.

Boivin, A., and Mesrobeanu, L., 1935 Recherches sur les antigenes somatiques et sur les endotoxines des bacte- ries. I Considerations generales et expose des tech­ niques utilisees. Rev. d. Immunol., _1, 553-569.

Boivin, A., and Mesrobeanu, L., 1936 Recherches sur les antigenes somatiques et sur les endotoxines des bacte- ries. II L*antigene somatiques complet (antigene 0) de certaines bacteries est le constituant principal de leur endotoxine. Rev. d. Immunol., 2, 113-144.

Bolven, A., and Mesrobeanu, L., 1937 Recherches sur les antigenes somatiques et sur les endotoxines des bacte­ ries. Ill Antigene soraatique 0 "complet" et variations bacteriennes. Etev. Immunol., 319-335-

Boivin, A., and Mesrobeanu, L., 1938 Recherches sur les antigenes somatiques et sur les endotoxines des bact­ eries. VI Sur la nature chimique et sur les proprietes biologlque des antigenes 0 et Vi du bacille typhique. Rev. d. Immunol., 4, 469-480.

73 74

Boyd, W.C., and Reeves, E., 1961 Specific inhibition of anti-D antibody by colominic acid. Nature, 1 9 1, 511.

Breed, R.S., Murray, E.G.D., and Smith, N.R., 1957 Bergey's Manual of Determinative Bacteriology, 7th ed., Williams" and Wilkins Co., Baltimore.

Crumpton, M.J., and Davies, D.A.L., 1958 The isolation of D-fucosamine from the specific polysaccharide of Chromobacterium violaceum (NCTC 7917). Biochem. J.. ET'729-736.------Davies, D.A.L., i960 Polysaccharides of Gram-negative bacteria. Adv. Carb. Chem., 1£, 271-340.

Davies, D.A.L., Morgan, W.T.J., and Mosimann, W., 1954 Preparation and properties of the "O'* somatic antigen of Shigella dysenteriae (Shiga). Biochem. J., 5 6, 572-581. —

Davies, D.A.L., Morgan, W.T.J., and Record, B.R., 1955 The specific polysaccharide of the dominant "0" somatic antigen of Shigella dysenteriae. Biochem. J., 60, 290-303.

Demos, F.L., and Randles, C.I., 1959 Unpublished data.

Dische, Z., 1947 A new specific color reaction of hexuronic acid. J. Biol. Chem., 1 6 7, 189-198.

Dische, Z., 1947a A specific color reaction for glucuronic acid. J. Biol. Chem., 171, 725-630.

Dische, Z., 1949 Spectrophotometric method for the determi­ nation of free pentose and pentose in nucleotides. J. Biol. Chem., 181, 3 7 9-3 9 2.

Dische, Z., 1950 A modification of the carbazole reaction of hexuronic acids for the study of polyuronides. J. Biol. Chem., 1 8 3, 489-494.

Dische, Z., 1955 New color reactions for determination of sugars in polysaccharides. Meth. Biochem. Anal., 2, 313-353.

Dische, Z., and Shettles, L.B., 1948 A specific color reaction of methylpentoses and a spectrophotometric micromethod for their determination. J. Biol. Chem., 175, 595-603. 75 Dische, Z., and Borenfreund, S., 1950 A spectrophotometric method for the microdetermination of hexosamines. J. Biol. Chem., 184, 517-522.

Dodd, M.C., Blgley, N.J.. and Geyer, V.B., i960 Specific inhibition of Rho(D) antibody by sialic acids. Science, 132, No. 3^37, 1398-1399. Evans, T.H., and Hibbert, H., 1946 Bacterial polysaccha­ rides. Adv. Carb. Chem., 2, 203-233.

Gardell, S., 1958 The determination of hexosamines. Meth. Biochem. Anal., 6, 289-317*

Gottschalk, A., i960 Sialic Acids: The Chemistry and Biology of Sialic Acids and Related Substances. Cambridge University Press, Cambridge, England.

Gyorgy, P., 1958 N-containing saccharides in human milk. In Chemistry and Biology of Mucopolysaccharides. Ciba Foundation Symposium, pp. i4o-15o. Edited by G.E.W. Wolstenholme and M. O ’Connor. J. and A. Churchill Ltd., London.

Hawk, P.B., Oser, B.L., and Summerson, W.H., 1954 Practical Physiological Chemistry, 13th ed., McGraw Hill7 Mew York.

Heidelberger, M., 1956 Chemical constitution and immunologi­ cal specificity. Ann. Rev. Biochem., 25, 641-658.

Heidelberger, M., 1959 In Proceedings Fourth International Congress of Biochemistry, Vienna, 195B. 1 52-70, Pergamon Press, London.

Heidelberger, M., Goebel, W.F., and Avery, O.T., 1925 The soluble specific substance of Friedlander*s bacillus, (three papers). J. Exper. Med. , 42, 701-745.

Hrubant, G.R., and Randles, C.I., 1957 Unpublished data.

Ikawa, M., and Niemann, C., 1949 A spectrophotometric study of the behaviour of carbohydrates in seventy-nine per cent sulfuric acid. J. Biol. Chem., 180, 923-931.

Ivanovics, G., and Bruckner, V., 1937 Die chemische strucktor der Kapselsubstanz der Milzbrandbazillus und der sero- logisch identischen spezifischen Substanz des Bacillus mesenterlcus. Z. Imraunitatsforsch., 9 0, 304-3TSTI 76 Kabat, E.A., 1958 Immunochemical approaches to polysaccha­ ride and mucopolysaccharide structure. In Chemistry and Biology of Mucopolysaccharides. Clba Foundation Symposium, pp. *»2-t>3. Edited by G. E. W. Wolstenholme and M. O'Connor. J. and A. Churchill Ltd., London.

Kabat, E.A., and Mayer, M.M., 1961 Experimental Immuno- chemistry, 2nd ed., Charles C. Thomas, Springfield, Illinois.

Kundlg, W., Neukom, H., and Deuel, H., 1961 Untersuchungen uber Getreideschleimstoffe. I Chromatographisch Frak- tionlerung von wasserloslichen Weizenmehl-pentosanen an Dlaethylamlnoaethyl-Cellulose. Helv. Chim. Acta., 44, 823-8 2 9.

Mehl, J.W., Pacovska, E., and Winzler, R.J., 1949 The amount of copper bound by protein In the Biuret reac­ tion. J. Biol. Chem., 177, 13-21.

Morgan, W.T.J., 1937 The isolation and properties of a specific antigenic substance from B. dysenteriae (Shiga). Biochem. J., J_l, 2003-2021.

Morgan, W.T.J., 1944 The chemical nature of bacterial antigens. Brit. Med. Bull., 2, 281-284.

Morgan, W.T.J., and Partridge, S.M., 1940 The fraction­ ation and nature of antigenic material isolated from Bact. dysenteriae (Shiga). Biochem. J., 34, I6 9-I9I.

Neeley, W.B., i960 Dextran: structure and synthesis. Adv. Carb. Chem., 15, 341-369.

Park, J.T., and Johnson, M.J., 1949 A submicrodetermination of glucose. J. Biol. Chem., l8l, 149-151.

Randles, C.I., 1957 In The Acid Mine-Drainage Problem in Ohio. Edited by E. Moulton. Eng. Exp. Stat., The 75fiTo State Univ. Bull. 166, 26, No. 5, 2 9.

Reissig, J.C., Strominger, J.L., and Leloir, L.F., 1955 A modified colorimetric method for the estimation of N-acetyl amino sugars. J. Biol. Chem., 217, 959-966.

Rosenberg, A., Howe, C., and Chargraff, E., 1956 Inhibition of Influenza virus haemagglutlnation by a brain lipid fraction. Nature, 177, 234-235.

Salton, M.R.J., i960 Microbial Cell Walls. John Wiley and Sons, Inc., New York. 77 Sehiff, P., 1934 Zur Kenntnis der Blutantigene des Shiga- bacillus. Z. Immunitatsforsch., 8 2, 46-52.

Seibert, F.B., and Atno, J., 1946 Determination of poly­ saccharide in serum. J. Biol. Chem. 163. 511-522.

Smith, I., 1958 Chromatographic Techniques. Interscience Publ., Inc.,'Wew York.

Somogyi, M., 19^5 A new reagent for the determination of sugars. J. Biol. Chem., 160, 61-68.

Springer, G.F., Williamson, P., and Brandes, W.C., 1961 Blocd-group activity of Gram-negative bacteria. J. Exp. Med., 113, 1077-1093.

Stacey, M., 1946 The chemistry of mucopolysaccharides and mucoproteins. Adv. Carb. Chem., 2, 161-201.

Stoffyn, P.J., and Jeanloz, R.W., 1954. Identification of amino sugars by paper chromatography. Arch. Biochem. Biophys., 5 2, 373-379-

Temple, K.L., and Koehler, W.A., 1954. Drainage from Bltu- minous Coal Mines. Eng. Exp. Stat., West Virginia liniv. Bull. 2 5, Series 54, No. 4-1, 15-16.

Trevelyan, W.E., Proctor, D.P., and Harrison, J.S., 1950 Detection of sugars on paper chromatograms. Nature, 166, 444-445.

Webster, M.E., Landy, M., and Freeman, M.E., 1952 Studies on Vi antigen. II Purification of Vi antigen from Escherichia coli 5396/38. J. Immunol., 6£, 135-142.

Westphal, 0., Peier, H., Luderitz, 0., and Fromme, I., 1954 Die Utosetzung und Charakterisierung von Zuckern mit Solfonylhydraziden. Biochem. Z., 326, 139-149.

Westphal, 0., Luderitz, 0., Eichenberger, E., and Neter, E., 1958 Mucopolysaccharides of Gram-negative bacteria: newer chemical and biological aspects. In Chemistry and Biology of Mucopolysaccharides. Ciba Foundation Symposium! ip. 107-199• Isdited by G. E. W. Wolsten­ holme and M. O'Connor. J. and A. Churchill, Ltd., London.

White, B., 1938 The Biology of the Pneumococcus. Hie Commonwealth Pund, New York. 78

Wilkinson, J.F., 1958 The extracellular polysaccharides of bacteria. Bacterid* Rev., 22, *16-73.

Winzler, R.J., 1958 Glycoproteins of Plasma. In ChemiBtry and Biology of Mucopolysaccharides. Ciba Foundation Symposium, pp. ^4 5-26'/. Edited by G.E.W. Wolstenholrae and M. O'Connor. J. and A. Churchill, Ltd., London. AUTOBIOGRAPHY

I, Alvin Gordon Lazen, was born in Baltimore, M&ryland

May 28, 1935* I received my secondary school education at

the Baltimore City College High School. My undergraduate

training was obtained at the University of Maryland, which granted me the Bachelor of Science degree in June, 1957*

I entered the Ohio State University in September, 1957 and served as a Graduate Assistant and Research Assistant until

August, 1958 when I entered the army. With the Army Chemi­ cal Corps at the Army Chemical Center, Maryland, I worked on a microbiological research problem from August, 1958 to

August, i9 6 0. In September, i9 6 0, I re-entered the Ohio

State University and began the research on which my disser­ tation is based. During the period I served as a Graduate

Teaching Assistant, Research Assistant, and for the final two years as Muellhaupt Fellow of the Graduate School.

79